How to use
- Click into whichever field matches the notation you already have — Binary, Octal, Decimal, Hexadecimal, or the Custom base row — and start typing. There is no convert button; the other four fields fill in as you type.
- For an unusual radix, set the number spinner beside the Custom base field to anything from 2 to 36, then read the value there or type into it directly.
- Paste without cleaning up first:
0x,0oand0bprefixes are accepted in every field, and separators such as spaces, underscores and commas are ignored on input. - Tick Group digits to break long outputs into readable chunks, or Show base prefixes to stamp
0x/0o/0bonto the emitted values. - Use Copy above any field to lift that one representation, or Clear to empty every field and refocus the binary box.
How it works
Positional notation gives each digit a weight that is a power of the base, so reading a number means multiplying every digit by its place value and adding the results. Rather than track exponents, the tool folds that sum into a single left-to-right accumulator, which is both faster and exact for arbitrarily long input.
Take the hexadecimal value b1e. Letter digits carry the values a=10 through f=15, so here b=11, 1=1 and e=14. The accumulator begins at 0 and, for each digit, multiplies the running total by the base (16) and adds the new digit: 0 → 0×16+11 = 11 → 11×16+1 = 177 → 177×16+14 = 2846. That single canonical number, decimal 2846, is what every other field renders from.
Rendering is the reverse: repeatedly divide the value by the target base and read the remainders from the bottom up. For 2846 that produces 101100011110 in binary (2048 + 512 + 256 + 16 + 8 + 4 + 2), 5436 in octal, and — if you set the custom base to 30 — 34q, where the trailing q is the digit worth 26. Change any one field and the accumulator re-derives from scratch, so the five views never drift out of sync.
Use cases & limitations
A hex dump, an RGB swatch, a Unix file mode, a subnet mask — much of computing is one integer wearing different clothes, and swapping those outfits is the whole job here. Firmware and embedded work lean on binary and hex constantly, and a quick “is 0x1f really 31?” check is faster in five live fields than in a mental calculation. When the number is actually a colour like #b1e, the colour converter understands its channels; when it is a network mask, the subnet calculator treats it as an address rather than a bare value.
The main limitation is that it converts whole numbers only. Fractions have no field: you cannot ask what 0.1 in decimal is in binary (a question with an endlessly repeating answer anyway), and there is no floating-point or fixed-point mode. Each value is also handled as a pure signed integer, not as a fixed-width register, so it will not draw you the 8-bit two’s-complement bit pattern behind a negative number. If your goal is inspecting bytes rather than numbers — an encoded payload, say — reach for the Base64 encoder/decoder instead.